ABOUT CHARACTERIZATION OF SURFACTANTS OUTSIDE THE HLB-SYSTEM

Ingegard Johansson, Akzo Nobel Surface Chemistry AB, SwedenIlja Voets, Wageningen University and Research Center, The Netherlands

ABSTRACT

In our work with new surfactants, having structures that are not easily treated with the Griffin orDavies HLB formulas, a need for characterization that would work for all kinds of structures has comeup. In this paper a possible system is discussed using a titration method in an emulsion system to findthe phase inversion, i.e. how far the structure of a specific surfactant is from forming aggregateshaving a spontaneous curvature of zero.In essence, the investigation can be characterized as a way of making an HLB scan using a mixture ofoil, water and surfactant and register the response to a continuous slow addition of a co-surfactant.The easiest response to measure is the conductivity of the stirred mixture, which reveals when themixture switches from an oil-continuous to a water-continuous emulsion or vice versa depending onthe starting point.

We have studied a model system of nonionic surfactants with straight and branched hydrophobes aswell as broad and narrow range ethylene oxide distribution in the hydrophilic part. The co-surfactanthas been Span® 80, a hydrophobic oleyl monoglyceride. Attempts were thus made to relate theconventional HLB:s to the data measured and to assign relative HLB:s to for instance the surfactantswith branched hydrophobes. Cloud points, CMC, surface tension, wetting, interfacial tension, foamand other functional properties were measured and the co-variation of these data has been analyzed.The system was stretched to cover also other surfactant types, for instance the sugar based nonionics.

From the titration the following conclusions could be drawn.- The broad range ethoxylates are less hydrophobic than the narrow range ones.- The ethoxylates with > 3EO:s made from branched hydrophobes require less co-surfactant,

i.e. act as more hydrophobic, than those with a straight chain in an emulsion fromdecane/water 50/50. This can be understood from the bulkiness of the hydrophobe making theform of the surfactant closer to a cylinder i.e. having a Critical Packing Parameter closer to 1,thus needing less hydrophobic additive to come to a zero curvature of the interface.

- Alkyl glucosides made from the same fatty alcohols show the same difference between thestraight and branched structures but to a larger extent.

- Comparing the EO-hydrophile with the glucosidic hydrophile does not give an unambiguouscorrespondence but depends on the balance in the molecular mixture. Both glucosidescontain the same average amount of glucose, 1.6, per hydrophobe. For the straight chain theglucosidic part corresponds approximately to 7 EO:s (isomerically pure) but for the branchedC10 it is more similar to 5 EO:s (narrow range).

INTRODUCTION

When oil and water are mixed, an emulsion is formed. In many applications the simultaneouspresence of a polar and an unpolar solvent is desirable. The resulting emulsion often needs to bestable. This stability is achieved by the addition of one or several emulsifiers. The problem ofchoosing the best system for this purpose is ubiquitous. Solutions have been sought in describing thestructure of the emulsifier in terms of HLB, Hydrophilic-Lipophilic Balance (1), or CPP, Critical PackingParameter (2). These attempts have not taken into account the influence of oil, temperature, salt etcwhich play important roles for emulsion stability. Shinoda (3) works with temperature effects and usesPIT (phase inversion temperature) to describe the function of emulsifiers, mainly of ethoxylatednonionic surfactants. Lately a systematic approach to use the PIT has been described (4) under theabbreviation CAPICO (Calculation of Phase Inversion in Concentrated Emulsions).

Typical for an emulsion, that is heated or cooled past the phase inversion temperature, is that itchanges its morphology. i.e. inverts by going from an O/W to a W/O emulsion or vice versa.Expressed in another way this means that the spontaneous curvature, Ho, of the interfacial filmchanges sign and passes a flat form where Ho = 0. This point is the balanced point or optimal pointwhere a bicontinuous microemulsion is formed and the interfacial tension is at its minimum.

Davies writes “ The spontaneous mean curvature Ho, determines whether the interfacial film wants tocurve onto its oil or its water side or prefers to be flat” (1).That this tendency plays an important role to decide if an O/W or a W/O emulsion will be formed, eventhough the curvature of an emulsion droplet is much less than that of a micelle, has been re-emphasized by Kabalnov and Wennerström (5) in the late nineties.

It has also been shown by Salager et al (6) that the most stable emulsions are to be found at a specificdistance from the balanced situation of Ho= 0, or in his therminology, SAD = 0 (surfactant affinitydifference).

Thus it is important to find this balanced point for the actual emulsion or formulation that you areoptimizing.

The aim of the work reported here is to find a practical way of deciding the balanced or optimal pointfor a mixture that can be practiced with all types of emulsifiers, also the ones that don’t react totemperature or salt. An automatic titration procedure is used by which a co-surfactant is added to anoil-water-surfactant blend under stirring at constant temperature. The morphology changes of theemulsion is followed with conductivity which suddenly goes to zero when the emulsion inverts fromO/W to W/O. By using reference compounds like isomerically pure ethoxylates and the same co-surfactant and oil the behavior of any other surfactant can be compared to systems with known HLB:sor whatever traditional description concepts that are familiar to the user.

Series of C10 nonionics with branched or straight hydrophobe and broad or narrow ethylene oxidedistribution or glucose as hydrophile, have been studied and compared. Physico-chemical data aswell as Optimal balance with the titration method have been decided. As co-surfactant Span 80®, oleylmonoglyceride , has been used.

EXPERIMENTAL

Products:Lab. products from Akzo Nobel Surface Chemistry AB:Guerbet C10 alcohol with 3, 5, 8 and 10 ethylene oxide units, both narrow and broad range distributionStraight C10 alcohol with 3, 5, 8 and 10 ethylene oxide units, both narrow and broad range distributionGlucosides from the same two alcohols with an average degree of polymerization of 1.6

Commercial products from Akzo Nobel Surface Chemistry AB: Berol ox 91-6 and 91-8 which are C9-11 alcohol with 6 or 8 ethylene oxide units, broad range.

Isomerically pure C10E5, C10E7, C10E8 from Sigma Aldrich and Fluka.Span® 80, oleyl monoglyceride, from Fluka, n-decane (>95%) from Merck. Water was distilled and de-ionized.

Critical Micelle Concentration.Surface tension was measured on a KSV unit with a Sigma 70 program using the Du Noüy ringmethod.

Contact angles.The contact angle was measured on Parafilm® with a FTÅ200 instrument in a climate controlled room(T=21±1°C, 45 ± 10%rH).

Cloud points.Mixtures of 1.0w% surfactant in distilled water, aqueous NaCl (2.0g/l; 0.034M) or 11% BDG wereprepared. The mixtures that were clear at room temperature were heated in a water bath until the

mixture became turbid. The mixtures that were turbid at room temperature were placed in ice until themixture became clear (or no difference in appearance could be determined in the 0-100ºC domain). Inboth cases cloud points were determined upon cooling.

FoamThe foam height is measured in a winding equipment with fixed measuring cylinders (500 ml) with 200ml of surfactant solutions in each, which is turned around 40 times during one minute. The foam heightis recorded immediately and after 1 min.

Conductivity measurements.N-decane (24.25g; 0.15mole), aqueous NaCl (24.25g; 2g NaCl/l; 0.034M) and surfactant (1.50g) werehomogenized in a thermostated glass vessel. Temperatures at the inversion point were in the region24.6ºC T 26.3ºC . The apparatus used to homogenize the mixture in the vessel was a stirringdevice (brand unknown) at maximum stirring speed and an Ultra Turrax (Polytron PT 1200) at astirring speed >15000rpm (level 5). Conductivity was measured with a Pt electrode (Metrohm 712conductometer) and temperature was measured with a thermometer (Physitemp BAT-10) while amixture of Span® 80/n-decane (8/3 w/w) was titrated in using a dosimat (Metrohm 665) at a dosingrate (dr) of 0.15ml/min. This mixture will be referred to as ‘cosurfactant’. Figure 11 shows theexperimental setup of the automatic titration technique.Some measurements were performed on the Scanalys® equipment. It is essentially the sameexperimental setup (www.scanalys.com), but one single stirrer with built-in electrode was used at 225and 300 rpm. The Span® 80/n-decane (8/3 w/w) was titrated in with a dosing rate of 4.44_l/s.

RESULTS AND DISCUSSION

Physico-chemical measurements.

Distributions.Typical ethylene oxide distributions for C10E5s are shown in Figures 1 and 2. It should be noted thatthere is a general difference in distribution between the two alcohols due to the sterically crowdedstructure of the branched decyl alcohol. This gives rise to more starting material being left in theethoxylated product which influences the distribution of the different parts in the mixture between theoil and the water phase and thus the physico-chemical behavior.

0

5

10

15

20

25

30

0 2 4 6 8 10 12 14 16 18 20

ethylene oxide units

are

a p

erc

en

tag

e

1D 5EO narr

1D 5EO broad

0

2

4

6

8

10

12

14

16

18

20

0 2 4 6 8 10 12 14 16 18 20

ethylene oxide units

are

a p

erc

en

tag

e

BrC10 5EO narr

BrC10 5EO broad

FIGURE 1. Ethylene oxide distributions of decyl FIGURE 2. Ethylene oxide distributions of branchedalcohol ethoxylates, broad and narrow. decyl alcohol ethoxylates, broad and narrow.

20

30

40

50

60

70

80

0,000 0,001 0,010 0,100 1,000 10,000 100,000

concentration (g/l)

su

rfa

ce

te

ns

ion

(m

N/m

)

1D 3EO

1D 5EO

1D 8EO

1D 10EO

FIGURE 3. Surface tension as a function of surfactant concentration of narrow range 1-decanolethoxylates (T=21± 1°C).

CMC.In Figure 3 typical CMC curves are given for the narrow range n-decyl alcohol series. In the case ofC10E3 not the whole product mixture is dissolved, which means that the value of the CMC refers to apart of the mixture and is not valid. It is consequently not included in the CMC discussion below.

1,00E-04

1,10E-03

2,10E-03

3,10E-03

4,10E-03

5,10E-03

6,10E-03

7,10E-03

0 2 4 6 8 10 12

degree of polymerization

cri

tic

al

mic

ell

e c

on

ce

ntr

ati

on

(m

ole

/l)

BrC10 broad

BrC10 narr

1D broad

1D narr

Berol OX 91

BrC10 glucoside

1D glucoside

FIGURE 4. Critical micelle concentrations of technical grade alcohol ethoxylates and alkylpolyglucosides as a function of degree of polymerization of the head group (T=21± 1°C).

In Figure 4 the dependence of the CMC:s on the chemical structures is shown. As expected (7), theCMC increases for the same type of surfactant with increasing <m> indicative of the increasinghydrophilicity of the surfactant monomer. Moreover, the steric repulsion between the surfactant headgroups that form the micelles increases with increasing <m> (average length of the EO chain)

Generally the branched products have higher CMC:s than the straight chain ones. The firstexplanation that comes to mind is the packing parameter (CPP). The packing parameter of surfactantswith straight chain hydrophobes is more favorable for the formation of (spherical) micelles than is thecase for branched hydrophobes. There is also a small difference between narrow and broad distributions. The broad range surfactantshave slightly more freedom in packing the different homologues into the micelles resulting in slightlylower critical micelle concentrations.

For comparison two commercial products based on a C9-11 alcohol blend are shown, which both havelower CMC:s than the pure C10 types due to the heterodispersity of the hydrophobe, again givingmore freedom in packing the different homologues into the micelles.

The glucosides show a huge difference between the two alcohols, again with a much higher CMC forthe branched product. It is well known that the hydrophilic/lipophilic balance of the alkyl glucosides iseven more sensitive to changes in the length of the hydrophobic chain than that of the ethoxylates (8).

Area per molecule at the air/water interfaceSome further information can be extrapolated from the surface tension versus log concentration plots.The area per molecule (ao ) at the air/ water interface as a function of <m> is shown in Figure 5.

Clearly, a0 increased with increasing length of the EO chain for all types of surfactants except for thebroad range Guerbet alcohol compounds. A similar reasoning as for the critical micelle concentrationis valid here. Due to the increase in hydration of the head group with increasing length of the EOchain, there is less free energy to be gained upon transfer of the surfactant from the bulk phase to theair-water interface. Moreover, larger head groups simply need more space. Therefore, with increasing<m> the equilibrium between surfactants in the bulk and at the interface shifts towards the bulk phase.

30

35

40

45

50

55

60

65

70

0 2 4 6 8 10 12

Ethylene oxide units

Are

a/h

grp

(Å

2/m

ole

cu

le)

Berol OX

BrC10 broad

BrC10 narr

1D broad

1D narr

FIGURE 5. Area per molecule at the air/water interface as a function of the average number ofethylene oxide units

Both branching of the hydrophobe and broadening of the EO distribution had an influence on theadsorption. Surfactants with the same type of hydrophobe showed similar area per molecule at theair/water interface for small head groups. Surfactants with the same type of ethylene oxide distributionshowed similar a0 values for bigger head groups. Thus the influence of the hydrophobe wasdominating for smaller head groups and the hetero dispersity of the EO chain was the dominant factorfor larger head groups. This reflects the increase in the size of the head group relative to thehydrophobe with increasing <m>. Because branched C10 hydrophobes are bulkier than straight C10hydrophobes, branching had an adverse effect on the adsorption at the air-water interface. It isobvious that broad range surfactants had smaller areas per molecule than narrow range surfactants.Most probably this is a result of the packing advantages of a broader EO distribution.

Surface tension at CMC

Another information to be gained from the CMC curves is the surface tension at or above CMC, seeFigure 6.

20

22

24

26

28

30

32

34

36

38

0 2 4 6 8 10 12

ethylene oxide units

su

rface t

en

sio

n (

mN

/m)

BrC10 broad

BrC10 narr

1D broad

1D narr

FIGURE 6. Surface tension at the critical micelle concentration as a function of the average number ofethylene oxide units..Surface tension increased with increasing <m> (except for the broad range Guerbet alcohol basedcompounds). For the same type of surfactant, the smaller the head group, the more surfactant adsorbsat the interface, and the lower the surface tension. Note also that the same inconsistency is found forthe branched C10 broad range compounds in the area/molecule graph.The Guerbet alcohol based surfactants showed lower surface tension values in spite of the relativelylow adsorption when compared to the straight chain surfactants. Apparently, the branched surfactantsform a more hydrophobic surfactant layer at the air-water interface. One plausible explanation is thefollowing. The branch may be positioned planar to the interface, while the rest of the surfactant tails ispositioned lateral to the interface. Therefore, the branch reduces the contact between water and airresulting in lower surface tensions for branched surfactants. However surfactant anchoring/geometryat the interface is possibly not (most) important in this respect, as the surface tension of neat GuerbetC10 alcohol is by itself lower than that of 1-decanol (as is the contact angle on Parafilm®).

Contact angles

The change of the contact angle of a droplet of 0.25% surfactant solution on Parafilm® (as ahydrophobic model surface) was followed with time with a high speed video camera. The values after60 s are shown in Figure 7 versus amounts of EO.

20

25

30

35

40

45

50

55

60

65

0 2 4 6 8 10 12

ethylene oxide units

co

nta

ct

an

gle

(°)

BrC10 broad

BrC10 narr

1D broad

1D narr

Berol OX 91

C10-Guerbet alcohol

1-decanol

water

FIGURE 7. Contact angle (_) on Parafilm® (60s after drop formation) as a function of the averagenumber of ethylene oxide units (<m>) of 0.25w% solutions of alcohol ethoxylates with <m> 5(T=21±1°C, 45±10%rH).

Narrow range 1-decanol based surfactants showed a larger contact angle than similar Guerbet C10alcohol based surfactants. This observation corroborates well with the observed trend in the surfacetension at the CMC. It is also in line with the fact that the contact angle of neat C10 Guerbet alcohol onParafilm® is smaller than the contact angle of neat 1-decanol.Broad range surfactants showed smaller contact angles than similar narrow range surfactants, whichagain can be explained by the adsorption being tighter at the air-water interface.

Cloud points

To fit the cloud points of all surfactants studied in this work into the experimental window, cloud pointsin 11% BDG (Figure 8) were determined.Cloud points increase with increasing average EO chain length. As the head groups are at the micellarsurface, micelles with larger head groups will appear more hydrophilic. Moreover, steric repulsionbetween the micelles will increase with increasing size of the head group.

Branched surfactants are less tightly packed in micelles. Therefore there is less steric repulsionbetween the micelles and the branched surfactants form a separate surfactant phase earlier. Anotherpossible explanation is the formation of a different shape and size of micelles for branched surfactants.These aggregates might have stronger intermicellar attractive forces resulting in lower cloud points.

30

40

50

60

70

80

90

0 2 4 6 8 10 12

ethylene oxide units

clo

ud

po

int

(°C

)

BrC10 broad

BrC10 narr

1D broad

1D narr

FIGURE 8. Cloud point (CP) as a function of the average number of ethylene oxide units <m> of1.0w% solutions of alcohol ethoxylates with <m> 5 in 11% butyl diethylene glycol.

Foam

0

20

40

60

80

100

120

140

160

180

200

BrC10 broad,0.1% BrC10 narr,0.1% C10 broad, 0.1% C10 narr, 0.1% BrC10 gluc,

0.05%

C10 gluc, 0.05% BrC10 gluc,0.28%

mm

20C 0 min

20C 1min

50C 0 min

50C 1min

FIGURE 9. Foam height, winding equipment, for different C10 alcohols with 5 EO compared to alkylglucosides from the same alcohols, immediately and after 1 min, at 20 °C and 50 °C.

The foam was measured for the different C10E5s and the corresponding glucosides, see Figure 9.The branched alcohol based surfactants all show much less stable foam, probably due to lesselasticity of the foam lamella created from the less well packed branched surfactants.Temperature effects are seen in one case, C10 narrow range, where the higher temperature giveslower foam depending on that the solution has reached its cloud point at 50 °C.

Conclusions of the physico-chemical measurements

Branching of the hydrophobe Increases CMC Increases area per head group

But Decreases surface tension Decreases contact angle Decreases cloud point Decreases foam.

Inversion points of emulsions, titration of co-surfactant.

Some of the conclusions of the preceding session are counter intuitive, like higher CMC (indicatingmore hydrophilic compounds) and at the same time lower surface tension and contact angles (morehydrophobic compounds).

By finding out how far each surfactant mixture is from creating a spontaneous curvature equal tozero in a 50/50 oil water emulsion we may be able to describe the behavior in the interface better .

FIGURE 10. Schematic representation of a typical titration path (arrow) during the conductivity scanwith the automatic titration technique, = co-surfactant/total surfactant concentration (w%), = totalsurfactant concentration (w%).

This is done via a titration procedure that was described earlier. How the titration moves the systemover the oil/water/surfactant phase diagram, keeping the oil/water ratio constant at 50/50 is shown inFigure 10. By adding a co-surfactant the total surfactant concentration increases somewhat and theratio between the two changes towards a more and more hydrophobic surfactant mixture, until itbalances to a planar structure and then flips over changing the nature of the emulsion from beingwater continuous to being oil continuous.

Poly

tro

n P

T120

0

Mix

er

Surfactant injection

Conductometer

Thermometer

In

Out

To thermostatic bath

FIGURE 11. Experimental set-up for the co-surfactant titration followed with conductometry.Commercially available as the Multi Parameter Scanning (MPS) instrument from Scanalys,(www.scanalys.com), with possibilities to measure turbidity, pH, conductometry and viscositysimultaneouslyFigure 11 shows the experimental set-up and Figure 12 a typical conductivity scan with conductivityagainst added co-surfactant, Span80 in ml.

FIGURE 12. Conductivity scan for C10E3 broad range with Span-80/n-decane/NaCl/water.

This is then evaluated and recalculated to find the upper and the lower inversion points (IP) accordingto the procedure described in Figure 13.

Iplower

Ipupper

FIGURE 13. : Linear fits to the first and last part of the steep conductivity drop in a plot of conductivityas a function of the amount of added cosurfactant in a pseudo quinary SOW system of 1-decylglucoside/n-decane/water/NaCl/Span® 80 (T=25ºC, dr=0.15ml/min, ss>15000rpm) are shown. IPupper

is determined as the breakpoint of a horizontal line through the maximum in the conductivity curve anda linear fit to the first portion of the steep drop in conductivity. IPlower is the amount of added surfactantin milliliters at the breakpoint of a horizontal line through the minimum in conductivity and a linear fit tothe last portion of the steep drop in conductivity

In Figure 14 all the the inversion points for the four series of ethoxylated surfactants, the isomericallypure ones and the alkyl glucosides are gathered in terms of w % of Span80 of the total surfactantpackage against degree of polymerization.

FIGURE 14. Inversion points (IPupper and IPlower) as a function of the degree of polymerization oftechnical grade alcohol ethoxylates, technical grade alkyl glucosides and isomerically pure alcoholethoxylates with 5, 7 or 8 EO units. Error bars indicate the deviation of two individual measurementson the system from the average value.

Results from the titration indicate:- The broad range ethoxylates are less hydrophobic than the narrow range ones.- The ethoxylates with > 3EO:s made from branched hydrophobes require less co-surfactant,

i.e. act as more hydrophobic, than those with a straight chain. This can be understood fromthe bulkiness of the hydrophobe making the form of the surfactant closer to a cylinder i.e.

having a Critical Packing Parameter closer to 1, thus needing less hydrophobic additive tocome to a zero curvature of the interface.

- Alkyl glucosides made from the same fatty alcohols show the same difference between thestraight and branched structures but to a larger extent.

- Comparing the EO-hydrophile with the glucosidic hydrophile does not give an unambiguouscorrespondence but depends on the balance in the molecular mixture. Both glucosidescontain the same average amount of glucose, 1.6, per hydrophobe. For the straight chain theglucosidic part corresponds approximately to 7 EO:s (isomerically pure) but for the branchedC10 it is more similar to 5 EO:s (narrow range).

Finally the average of the inversion point values is taken as the inversion point and used for atentative coordination of an Equivalent HLB value comparing with the results for the isomerically pureC10E5, C10E7 and C10E8, for which the HLB:s can be calculated according to Griffin, see Figure 15.

FIGURE 15. : Extrapolation of a linear fit on the inversion points (average of upper and lower) as afunction of the HLB number (Griffin) of pure C10Em alcohol ethoxylates is shown. The equationobtained in this manner provides a way to determine the Equivalent HLB number of any (mixture of)surfactant(s) once its inversion point in the same system (i.e. the same temperature, pressure, type ofoil) is established.

Conclusions overall

Technical grade C10E<m> surfactants are blends of surfactant homologues with an average length ofthe EO distribution at <m>. Depending on catalyst the distribution of homologues can be made moreor less narrow.

Broad range surfactants show:• higher amounts of unreacted alcohol• higher critical micelle concentrations• smaller areas per molecule at the air/water interface• lower surface tension• lower contact angles• higher cloud points• higher inversion points

The behavior of technical grade C10E<m> surfactants is shown to be dependent upon both the shapeand mean of the EO distribution. It is also strongly dependent upon the surfactant environment.Therefore, it is not possible to universally classify broad range surfactants as more hydrophilic orhydrophobic than narrow range surfactants. It is not possible to determine an increment in (a factordescribing the) hydrophilicity due to head group heterodispersity as the increment is dependent upon<m> and surfactant environment (i.e., emulsion or aqueous solution). This results from the fact that itis not possible to treat the surfactant blend as a pure surfactant with a number of EO units equal to

<m> (so called lumping). Selective adsorption, partitioning and phase separation of low and/or highhomologues make lumping impossible.

Likewise, it is not possible to universally classify branched surfactants as more hydrophilic orhydrophobic than straight chain surfactants. Therefore, it is not possible to determine an increment in(a factor describing the) hydrophilicity due to branching as the increment is dependent upon <m> andsurfactant environment (i.e., emulsion or aqueous solution).

Branched surfactants show:• higher amounts of unreacted alcohol• higher critical micelle concentrations• larger areas per molecule at the air/water interface• lower surface tension• lower contact angle• lower cloud points• less foam• lower inversion points

Guerbet branching of the hydrophobe results in both hydrophobic and hydrophilic shifts in surfactantbehavior. This is partly due to the influence of the unreacted alcohol which decreases with increasing<m>. Moreover, the increased CPP due to Guerbet branching results in both hydrophilic shifts (higherCMC) and hydrophobic shifts (lower cloud point).

All studied alkyl polyglucosides and polyoxyethylene alkyl ethers spontaneously form O/W emulsions(3w% surfactant) in aqueous NaCl/n-decane (_=1.0 by weight) except narrow range 1D 3EO (W/Oemulsion). It is possible to invert emulsions with alkyl polyglucosides and C10E<m> for <m> 8 if theratio of cosurfactant to (total) surfactant necessary for the inversion is not too high (approximately 50%). Partitioning of surfactant homologues into the water and oil phases and surfactant geometry arethe dominant factors that influence the position of the balanced point.

With a simple automatic titration technique the ratio of hydrophobic co-surfactant to total surfactant (_)necessary to invert an emulsion from O/W to W/O, can be obtained. The HLB numbers (Griffin) ofisomerically pure straight alcohol ethoxylates were plotted as a function of (_). The equivalent HLBnumber (EHLB) of any (mixture of) surfactant(s) can be determined from this graph and thecorresponding equation once _ is measured with the automatic titration technique. This method ofsurfactant classification is more precise and refers to a specific oil/water/temperature system which theHLB number does not. In this respect it is more similar to the HLB temperature concept launched byShinoda (3).

REFERENCES

1. Davis, H. T., Colloids and Surfaces, A: Physicochemical and Engineering Aspects 91, 9-24(1994).

2. Israelachvili, J.N., Intermolecular and surface forces, Academic press ltd, London, (1992)3. Shinoda, K., Friberg, S., Emulsions and solubilization, Wiley, New York (1986).4. Wadle, A., Tesmann, H., Leonhard, M., Förster, T., in Surfactants in Cosmetics, ed. Rieger,

M.M., Rhein, L.D., Marcel Dekker Inc, New York, Basel, 207-224, (1997)5. Kabalnov, A., Wennerström, H., Langmuir 12, 276-292 (1996)6. Pérez, M., Zambrano, N., Ramirez, M., Tyrode, E., Salager, J., J. Dispersion Science and

Technology 23(1-3), 55-63 (2002)7. Rosen, M. J., Cohen, A. W., Dahanayake, M., Hua, X. Y., Journal of Physical Chemistry 86(4),

541-5 (1982).8. Nilsson, F., Söderman, O., Hansson, P., Johansson, I., Langmuir 14, 4050-4058 (1998)